Skip to main content

Lab 2.3 - Monitoring System Components

Objectives​

By the end of this lab, you will be able to:

  • Monitor Control Plane components
  • Monitor Worker Nodes
  • Analyze system metrics
  • Identify potential issues

Prerequisites​

  • A working Kubernetes cluster
  • kubectl installed
  • Metrics Server installed (for kubectl top)

Exercise 1: Monitoring the Control Plane​

Step 1: Component Status​

# View component status (deprecated but useful)
kubectl get componentstatuses

# View Control Plane Pods
kubectl get pods -n kube-system

Step 2: Component Logs​

# API Server logs
kubectl logs -n kube-system -l component=kube-apiserver --tail=100

# Controller Manager logs
kubectl logs -n kube-system -l component=kube-controller-manager --tail=100

# Scheduler logs
kubectl logs -n kube-system -l component=kube-scheduler --tail=100

Exercise 2: Monitoring Nodes​

Step 1: Node Health​

# View the status of all nodes
kubectl get nodes

# Node details
kubectl describe node <node-name>

# View node conditions
kubectl get node <node-name> -o jsonpath='{.status.conditions}'

Step 2: Node Metrics​

# Resource usage
kubectl top nodes

# Resource details
kubectl describe node <node-name> | grep -A 20 "Allocated resources"

Exercise 3: Events and Alerts​

Step 1: Recent Events​

# All events
kubectl get events --all-namespaces --sort-by='.lastTimestamp'

# Error events
kubectl get events --all-namespaces --field-selector type=Warning

# Events in a namespace
kubectl get events -n kube-system

Step 2: Component Events​

# Events related to a system Pod
kubectl describe pod <pod-name> -n kube-system | grep Events -A 10

Exercise 4: Cluster Health​

Step 1: General Check​

# Cluster information
kubectl cluster-info

# Version
kubectl version

# Available resources
kubectl api-resources

Step 2: System Namespace​

# All resources in kube-system
kubectl get all -n kube-system

# System secrets
kubectl get secrets -n kube-system

# System ConfigMaps
kubectl get configmaps -n kube-system

Exercise 5: Troubleshooting​

Step 1: Common Issues​

# Pods in error
kubectl get pods --all-namespaces --field-selector status.phase!=Running

# Pods restarting
kubectl get pods --all-namespaces | grep -E "Error|CrashLoopBackOff"

# Nodes NotReady
kubectl get nodes | grep NotReady

Step 2: Log Analysis​

# Logs of a crashing Pod
kubectl logs <pod-name> -n <namespace> --previous

# Logs of all containers
kubectl logs <pod-name> -n <namespace> --all-containers=true

Reflection Questions​

  1. Which components are critical for the cluster's operation?
  2. How do you detect a problem with a component?
  3. Which metrics are most important to monitor?
  4. How would you react if a node becomes NotReady?

Cleanup​

No cleanup necessary.

Next Steps​

Module 3: Pods and Deployments
Module 2 Quiz: Validate your knowledge

Lab created on: December 2024